Add OpenDevShop support

[omega8cc]

This was originally requested a long time ago, so let's make this happen!

Some related issues:

#439
#720
#473
https://www.drupal.org/node/1935810

[heliogabal]

+1, this would tremendously ease my current workflow.

[ahillio]

+++! Would love to have this on your hosted service!!

[radimklaska]

+1 It would be BIG win to have octopus instance running devshop!

[thebennos]

+1 for this

[arttus]

After working with both platforms for a bit i'd say combining OpenDevshop and BOA is not a practical idea. The best direction now for BOA would be better git support.

OpenDevshop currently has the best front-end for Aegir. AND you could get the Devshop Devmaster front end running on BOA. But we need to focus our energies, so does it makes sense for two projects with so much in common to spend extra time shoehorning Devsmasters frontend into BOA, OR would be a waste of effort?

As a long time user of BOA (2010) i'm torn, I am grateful for both and would like each project to succeed on their own terms... but i guess i have to say is what i'd like to see now is a BOA playbook for OpenDevshop.

What are all the best features of BOA, let's write an Ansible playbook.

[omega8cc]

@arttus -- We don't see any duplication/waste of effort here. Why not to support it like we support vanilla Aegir?

Not sure what you mean by "The best direction now for BOA would be better git support"? It is not related to BOA, but to Aegir, which happens to be the control panel for Drupal we support. Maybe Aegir upstream could improve git support, or embrace ideas from OpenDevShop?

OpenDevShop has a nice concept and front-end, but it doesn't provide any advanced system/server level features BOA provides, so both projects and their users/audience could only benefit from working together.

As for Ansible, it was discussed in the past. Basically, it is not the right tool for BOA (too limited) and there is no benefit in rather big effort needed to cut BOA monolithic bash spaghetti into pieces (modules and roles) Ansible expects. BOA could benefit from a fundamental rewrite, but why exactly should we translate it into Ansible playbooks/modules/roles? We should rather focus on converting as much as possible into deb packages, remove old cruft and overcomplicated workarounds you are not going to translate to Ansible language anyway. Then we could take a look again, if such idea still makes any practical sense.

[arttus]

@omega8cc yeah I just meant with everyone's workflow pretty much solidly based on git, if BOA had better git support it would be much easier to use. Not sure how much work this integration would be https://www.drupal.org/project/hosting_git.

Totally agree about OpenDevShop being more of a developer tool then a production environment.

If adding OpenDevshops DevMaster front-end to BOA is a relatively simple matter like supporting vanilla Aegir's front-end then that would be a very useful addition indeed.

I was also thinking about the fundamental difference between the two projects, and i wasn't suggesting cutting up BOA's scripts into a playbook as much as building a high performance playbook that was designed with all of best features of BOA.

If deb packages make more sense then Ansible, maybe that's a better way to go, I wouldn't know about that level of detail of BOA's internals.

[arttus]

I forgot to mention why designing a new playbook for BOA comes to mind, OpenDevshop already supports it.

ubuntu@devshop:~$ sudo bash install.sh --makefile=devmaster-BOA.make --playbook=playbook-BOA.yml